This proposal investigates the possibilities of using a robotic manipulator for stroke rehabilitation. Robots seem to be a logical choice for rehabilitation because they allow prolonged treatment sessions, they measure performance objectively,, and they can accurately exert many types of forces on the patient (force fields). Because the optimal method for applying these forces is not yet known, we specifically focus here on force fields that capitalize on the natural adaptive capacities of the nervous system. Research on healthy individuals has revealed that subjects show dramatic adaptation to a disturbing force field, and such adaptation occurs progressively without subjects being aware of the adaptive process. Most importantly, they exhibit characteristic after- effects when the disturbing force field is unexpectedly removed. Since 1) these after-effects can be predict by neuro-musculoskeletal models, 2) the ability to adapt appears to be preserved in many hemiparetics, and 3) stroke patients tend to show stereotypical errors in movement, we hypothesize that the appropriate force field training can yield desirable after-effects in hemiparetics. The main goal of this project is to perform a series of preliminary investigations that should help to determine whether such an approach is feasible for attaining desirable rehabilitation outcomes. We plan to use a programmable two-link robot that exerts forces during reaching movements in the horizontal plane. First, we plan to investigate to the magnitude of after effects on strike patients as compared to healthy individuals, determine how lesion site influences the ability to adapt under these conditions. Second, we plan to conduct two experiments that investigate various methods for force field adaptation as a form of therapy. The results of these experiments should lead to the development of more sophisticated robotic applications for rehabilitation and provide a guide for more long-term clinical trials.